Use of Natural Language Understanding to Facilitate Surgical De-Escalation of Axillary Staging in Patients With Breast Cancer.

PURPOSE: Natural language understanding (NLU) may be particularly well equipped for enhanced data capture from the electronic health record given its examination of both content-driven and context-driven extraction. METHODS: We developed and applied a NLU model to examine rates of pathological node positivity (pN+) and rates of lymphedema to determine whether omission of routine axillary staging could be extended to younger patients with estrogen receptor-positive (ER+)/cN0 disease. RESULTS: We found that rates of pN+ and arm lymphedema were similar between patients age 55-69 years and ≥70 years, with rates of lymphedema exceeding rates of pN+ for clinical stage T1c and smaller disease. CONCLUSION: Data from our NLU model suggest that omission of sentinel lymph node biopsy might be extended beyond Choosing Wisely recommendations, limited to those older than 70 years and to all postmenopausal women with early-stage ER+/cN0 disease. These data support the recently reported SOUND trial results and provide additional granularity to facilitate surgical de-escalation.

Identifying Genomic Alterations in Patients With Stage IV Breast Cancer Using MammaSeq: An International Collaborative Study.

BACKGROUND: Identification of genomic alterations present in cancer patients may aid in cancer diagnosis, prognosis and therapeutic target discovery. In this study, we aimed to identify clinically actionable variants present in stage IV breast cancer (BC) samples. MATERIALS AND METHODS: DNA was extracted from formalin-fixed paraffin-embedded samples of BC (n = 41). DNA was sequenced using MammaSeq, a BC-specific next-generation sequencing panel targeting 79 genes and 1369 mutations. Ion Torrent Suite 4.0 was used to make variant calls on the raw data, and the resulting single nucleotide variants were annotated using the CRAVAT toolkit. Single nucleotide variations (SNVs) were filtered to remove common polymorphisms and germline variants. CNVkit was employed to identify copy number variations (CNVs). The Precision Medicine Knowledgebase (PMKB) and OncoKB Precision Oncology Database were used to associate clinical significance with the identified variants. RESULTS: A total of 41 samples from Turkish patients with BC were sequenced (read depth of 94-13,340; median of 1529). These patients were diagnosed with various BC subtypes including invasive ductal carcinoma, invasive lobular carcinoma, apocrine BC, and micropapillary BC. In total, 59 different alterations (49 SNVs and 10 CNVs) were identified. From these, 8 alterations (3 CNVs - ERBB2, FGFR1, and AR copy number gains and 5 SNVs - IDH1.R132H, TP53.E204∗, PI3KCA.E545K, PI3KCA.H1047R, and PI3KCA.R88Q) were identified to have some clinical significance by PMKB and OncoKB. Moreover, the top 5 genes with the most SNVs included PIK3CA, TP53, MAP3K1, ATM, and NCOR1. Additionally, copy number gains and losses were found in ERBB2, GRB7, IGFR1, AR, FGFR1, MYC, and IKBKB, and BRCA2, RUNX1, and RB1, respectively. CONCLUSION: We identified 59 unique alterations in 38 genes in 41 stage IV BC tissue samples using MammaSeqTM. Eight of these alterations were found to have some clinical significance by OncoKB and PKMB. This study highlights the potential use of cancer specific next-generation sequencing panels in clinic to get better insight into the patient-specific genomic alterations.

Characterizing Molecular Variants and Clinical Utilization of Next-generation Sequencing in Advanced Breast Cancer.

Women with advanced breast carcinomas have few therapeutic options. Recent advances in genomic profiling represent a new paradigm of cancer classification and treatment, but experience with genomic testing in a clinical setting remains limited. We retrospectively determined the genomic variants and correlate these with histology [histomorphologic subtype, nuclear grade, standard immunohistochemistry (IHC)] and clinical utilization (ordering, turnaround time, report review, and targeted therapy). Among 48 patients, 2 showed no genetic alterations, 11 (23%) showed variants of unclear significance only and 35 (73%) showed variant(s) affecting function (VaF) and/or variants of unclear significance. Overall, 119 variants were observed in 20 of 50 tested genes. Each patient had a unique molecular profile, with numerous (n=58) variants not previously reported in breast cancer. VaF detected in more than 2 patients included: TP53 (n=21), PIK3CA (n=20), and FGFR1 (n=3). VaF comprised 46 single nucleotide variants (79%), 7 amplifications (12%), 3 frameshifts (5%), 1 insertion (2%), and 1 deletion (2%). The tested samples had very high Ki67 index (average 57%±23%) and approximately half were hormone receptor and HER2 negative (25/46, 54%). Metastatic breast carcinomas showed a higher average VaF versus breast-localized tumors (1.3±0.99 vs. 0.18±0.60, P<0.05). Next-generation sequencing reports were promptly reported and reviewed (average 1 to 2 d) and 7 (∼25%) of potentially eligible patients received targeted therapy. Advanced breast cancers show unique landscapes of genetic variants. Most testing was done in late disease, often in metastatic and receptor-negative carcinomas. Next-generation sequencing results were promptly reported and reviewed, but the utilization of targeted therapies was limited.